Deep Learning Holdup Solution From Neutron Capture And Inelastic Scattering

1. A method for determining a borehole holdup, the method comprising: logging a well with a pulsed-neutron logging (PNL) tool; receiving, via the PNL tool, transient decay measurements, capture spectrum measurements, and inelastic spectrum measurements; extracting information from each of the capture spectrum measurements, the inelastic spectrum measurements, and the transient decay measurements; combining at least some of the extracted information to form an input; inputting the input into a model; and determining the borehole holdup with the model.

2. The method of claim 1, wherein the extracting information from the transient decay measurements comprises extracting new features, physics, and decay data.

3. The method of claim 2, wherein the extracting information from each of the capture spectrum measurements and the inelastic spectrum measurements comprises extracting new features, physics, and an entire spectrum.

4. The method of claim 3, further comprising inputting auxiliary data into the into the artificial neural networks.

5. The method of claim 4, wherein the inputting the auxiliary data comprises inputting borehole size, casing size, tool standoff, and/or formation variations.

6. The method of claim 4, wherein the extracting new features comprises extracting indirect physical parameters.

7. The method of claim 1, wherein the extracting information from the transient decay measurements comprises extracting a time constant.

8. The method of claim 1, wherein the inputting the input comprises inputting all of the extracted information as a single spectral image.

9. The method of claim 1, further comprising extracting elemental yield from the capture spectrum measurements.

10. The method of claim 1, further comprising extracting elemental gamma ratios from the inelastic spectrum measurements.

11. A system for determining a borehole holdup, the system comprising: a pulsed-neutron logging (PNL) tool operable to measure capture spectrum, inelastic spectrum, and transient decay; and a computer operable to: extract information from each of capture spectrum measurements, inelastic spectrum measurements, and transient decay measurements; combine at least some of the extracted information to form an input; input the input into a model; and determine the borehole holdup with the artificial neural networks.

12. The system of claim 11, wherein the computer is further operable to extract new features, physics, and decay data from the transient decay measurements.

13. The system of claim 11, wherein the computer is further operable to extract new features, physics, and an entire spectrum from each of the capture spectrum measurements and the inelastic spectrum measurements.

14. The system of claim 11, wherein the computer is further operable to extract indirect physical parameters from each of the capture spectrum measurements, inelastic spectrum measurements, and transient decay measurements.

15. The system of claim 11, wherein the computer is further operable to input auxiliary data into the artificial neural networks.

16. The system of claim 15, wherein the auxiliary data comprises borehole size, casing size, tool standoff, and/or formation variations.

17. The system of claim 11, wherein the computer is further operable to extract a time constant from the transient decay measurements.

18. The system of claim 11, wherein the computer is further operable to input the input as a single spectral image into the artificial neural networks.

19. The system of claim 11, wherein the computer is further operable to extract elemental yield from the capture spectrum measurements.

20. The system of claim 11, wherein the computer is further operable to extract elemental gamma ratios from the inelastic spectrum measurements.